0

enter image description here

please, take me info, this bad or good?

I am real don't understand....

i'm know, also when

Mean Absolute Deviation (MAD): In [0,∞), the smaller the better

Root Mean Squared Error (RMSE): In [0,∞), the smaller the better

Median Absolute Error (MAE): In [0,∞), the smaller the better

Mean Squared Log Error (MSLE): In [0,∞), the smaller the better

but,

R², coefficient of determination: In (−∞,1] not necessarily the bigger the better

link topic info

https://datascience.stackexchange.com/questions/42760/mad-vs-rmse-vs-mae-vs-msle-vs-r²-when-to-use-which

enter image description here

enter image description here

Iakov Andreev
  • 1
  • 1
  • 1
    Welcome to Data Science! I have no idea why people think you should use MSE instead of $R^2$ when, at least depending on the calculation, $R^2$ is a function of MSE. Consequently, how is your $R^2$ being calculated? // While I believe myself to have a [good reason](https://stats.stackexchange.com/questions/590199/how-to-motivate-the-definition-of-r2-in-sklearn-metrics-r2-score) to disagree with a popular way of calculating out-of-sample $R^2$, [the relationship with MSE exists whether you do it my way or the way in SKLearn](https://stats.stackexchange.com/a/592424/247274). – Dave Nov 17 '22 at 03:35
  • Thank you answer, Dave! Yes, I also think that it is necessary to understand when a metric makes sense, not only on an intuitive level, but also to really understand what it does and what it means. Your suggestions are justified. But help me understand, that is, in this case, MCE has more weight, unlike P2? That is, in other words, I could interpret this result as excellent, if you do not take into account the evil metric P2? Please tell me Dave! And i am add new plot, you have idea about he? – Iakov Andreev Nov 17 '22 at 08:59
  • What are MCE and P2? – Dave Nov 17 '22 at 14:41
  • I'm sorry, I wrote at night, an error. This is -> MSE and R^2 – Iakov Andreev Nov 17 '22 at 17:58
  • How are you calculating the $R^2$ values you’ve posted? – Dave Nov 17 '22 at 18:04
  • pycaret library in Python – Iakov Andreev Nov 18 '22 at 12:35
  • How does it do the calculation, and what data do you input into the function? – Dave Nov 18 '22 at 12:36
  • https://scikit-learn.org/stable/modules/generated/sklearn.metrics.r2_score.html – Iakov Andreev Nov 18 '22 at 12:37
  • That calculation is related to MSE and RMSE, so any claim that “higher isn’t necessarily better” means that “lower isn’t necessarily better” for MSE or RMSE, which contradicts your references (correctly contradicts them, I’d say). With that in mind, what contradictions do you see in your chart of reported performance? – Dave Nov 18 '22 at 12:47

0 Answers0